Effect of very fine particles on workability and strength of concrete made with dune sand

Construction and Building Materials 47 (2013) 131–137

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Effect of very fine particles on workability and strength of concrete made with dune sand Fu Jia Luo a, Li He b, Zhu Pan a,⇑, Wen Hui Duan a, Xiao Ling Zhao a, Frank Collins a a b

Department of Civil Engineering, Monash University, Clayton, VIC 3800, Australia Institute for Frontier Materials, Deakin University, Waurn Ponds, VIC 3216, Australia

h i g h l i g h t s  Sand to cement ratio can affect properties of dune sand concrete (DSC).  Dune sand grain with size smaller than 175 lm (VFP) can affect the hydration.  VFP increases strength of DSC due to nucleation and pozzolanic effects.  DSC and river sand concrete have comparable engineering properties.  Australian dune sand can be used as fine aggregates for making concrete.

a r t i c l e

i n f o

Article history: Received 31 January 2013 Received in revised form 2 May 2013 Accepted 5 May 2013

Keywords: Very fine particles Strength Elastic modulus Workability Pozzolanic effect Nucleation effect

a b s t r a c t This paper presents the study on the properties of concrete made with dune sand from Australian desert. With constant water–cement ratio of 0.5, dune sand concrete (DSC) and the corresponding reference samples (concrete made with river sand) were prepared with sand–cement (S/C) ratio ranging from 0.91 to 2.28. In comparison to river sand, dune sand possesses a higher amount of very fine particles (VFPs) with grain size smaller than 175 lm. These VFPs are found to modify the properties of concrete by different mechanisms depending on the level of S/C ratio. At low level of S/C ratio (S/C < 1.41), VFPs can fill the porosities between cement pastes and aggregates and has no negative effect on workability and, the highest slump (105 mm) for DSC was found at S/C ratio of 1.18. Moreover, at low level of S/C ratio, the strength of DSC is comparable or even higher than that of river sand concrete (RSC); the higher strength of DSC can be attributed to the heterogeneous nucleation and pozzolanic effect brought by VFPs which enhances cement hydration. At high level of S/C ratio (S/C > 1.41), excessive VFPs absorb large quantities of water on their surface and lead to the reduction in workability for DSC. As a result, more air bubbles are introduced during compaction, leading to higher air content in DSC compared to RSC. The air bubbles increase porous space in cement paste and thereby reducing the strength of DSC. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Concrete manufacturing requires a large amount fine aggregate which is generally natural sand exploited from river channel and floodplain [1]. However, with the ever-increasing demand of aggregates due to the booming infrastructure development, river mining has led to serious environmental impacts, including dust arising, riverbank erosion, shifting of river course, suspendedsolids contamination and flooding, etc. [2]. Besides, the mining activities can also lead to the loss of coastal ecosystem, damage to infrastructure (e.g. roads and bridges) and potential destruction of tourism archaeological site, etc. [3]. Under the circumstances, the use of alternative materials for replacing river sand becomes ⇑ Corresponding author. Tel.: +61 0399051291 E-mail address: [email protected] (Z. Pan). 0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2013.05.005

vitally important for making concrete in an environmentally friendly manner. As one possible solution, the dune sand from desert regions is abundant in some parts of the world, especially some Africa and Middle East countries. In those regions, there are also problems associated with increasing shortage of coarse sands traditionally used in concrete. Therefore, if dune sand can be utilised, this situation will be somewhat ameliorated. Due to the environmental benefits, the topic on cementitious materials made with dune sand has aroused increasing interest. During the past decade, a number of researches have been conducted to study the characteristics of dune sand as well as the properties of cementitious materials incorporating it. Al Harthy et al. [4] have investigated the properties of concrete made with Sharkiya (Oman) dune sand which was partially introduced (10–100%) as fine aggregate. It was found that the optimum replacement ratio was around 50%, at which the concrete has its

132

F.J. Luo et al. / Construction and Building Materials 47 (2013) 131–137

best workability whilst the reduction in compressive strength was less than 25%. Zhang et al. [5] have investigated the complete use of dune sand as fine aggregate for making concrete. They have studied the performance of mortar and concrete made with Tenggeli (China) dune sand and attempted to improve the workability of dune sand concrete (DSC) by using superplasticizers. It was found that the superplasticizer could improve not only compressive strength but also workability of DSC. A recent research by Jin et al. [6] has explored the feasibility of using dune sand from Maowusu sandy land (China) in high-strength concrete and a compressive strength in excess of 65 MPa was reported. The above authors concluded that the dune sand can be used in making concrete for structural application. Besides, dune sand concrete also provides competent performance to other end-use applications. Wang et al. [7] and Qin et al. [8] have investigated the mechanical properties of concrete-filled tube made with dune sand (in China); results showed that the dune sand concrete-filled tube had higher flexural strength than that made with river sand. Khay et al. [9] have examined the feasibility of using dune sand concrete for pavement application. The sand was from Tunisian Sahara (Tunisia). It was found that the mixing of 60% dune sand and 40% crushed sand as fine aggregate produced competent concrete for pavement application. As discussed above, various properties of concrete made with dune sand from different desert regions have been investigated in previous studies and it was concluded that dune sand may provide a readily available alternative material for use as fine aggregate in concrete. However, the results from previous studies show a possible degradation in compressive strength when dune sand is completely used as fine aggregate in concrete [4,6]. This is generally attributed to the poor gradation of dune sand, which makes it different from normally-used sand for making concrete. In general, dune sand consists of a considerable amount of very fine particles (VFP). Typically 25% by weight of grains in dune sand are smaller than 150 lm; in contrast, this fraction is generally less than 6% as specified in ASTM C33 for fine aggregate used in concrete. It is interesting to note that VFP is also a major component in mineral admixtures such as silica fume and fly ash. These pozzolanic mineral admixtures have been used as a partial substitution for Portland cement for many years. The effects of these admixtures on properties of concrete have been extensively studied [10,11]. In general, these admixtures present binding activity because they can enhance cement hydration. Two major effects are observed on enhancement of cement hydration when these mineral admixtures are used in cementitious materials [12–15]. The first effect is heterogeneous nucleation which is a physical process leading to a chemical activation of the hydration of cement. It is related to the nucleation of hydrates on foreign mineral particles, which catalyses the nucleation process by reducing the energy barrier. Qualitatively, if the surface of the solid substrate matches well with the crystal, the interfacial energy between the two solids is smaller than the interfacial energy between the crystal and the solution, and nucleation may take place at a lower saturation ratio on a solid substrate surface than in pore solution without mineral admixture. The mineral powder used does not have to be reactive itself since its principal function is to provide nucleation sites for hydrates. This effect becomes significant for VFP as the decrease in particle size favours nucleation. The second effect is pozzolanic effect which is a chemical process leading to the increase in compressive strength. The mineral admixtures having pozzolanic activity will, in finely divided form and in the presence of water, react chemically with calcium hydroxide at ordinary temperature to form additional calcium silicate hydrates. These reaction productions fill in pores and result in a refining of the pore structure leading to improved strength and

durability. This effect is principally dependent on the fineness of materials and the amount of soluble amorphous silica in materials. A recent study [16] on cement paste containing dune sand powder has identified the occurrence of pozzolanic reaction. This is further confirmed by Alhozaimy et al. [17] that autoclaved curing can promote the pozzolanic reactivity of dune sand and therefore greatly increases the concrete strength. Therefore, when used to make concrete, dune sand should not be merely considered as inert fillers as normal fine aggregate but rather as an active component due to its pozzolanic reactivity and the heterogeneous nucleation effect. In this respect, the amount of VFP plays an important role in optimising the properties of concrete made with dune sand, which has received little attention. Therefore, the aim of this paper is to investigate the engineering properties of DSC with various VFP content. The investigated properties include air content, workability, compressive strength, tensile strength and elastic modulus. Also, it is worth mentioning that the dune sand used in this study is red dune sand from central Australia for the following reasons: (1) dune sand from different regions can exhibit quite different properties and there is no previous research on the use of Australia dune sand in concrete and (2) desert fields cover 40% of mainland Australia and are widely spread around the central continent, which provides abundant and easily available dune sand. Moreover, construction sites in central Australia are far from aggregates production quarries, making it uneconomical for aggregate transporting. Therefore, the results obtained in the current research also provide a guideline for the use of dune sand in concrete whenever suitable sand materials are not economically available in these areas.

2. Experimental program 2.1. Materials In this study, ASTM C150 Type I Ordinary Portland cement was used for all the concrete mixtures. The chemical compositions of cement were analysed by X-ray Fluorescence (XRF) as shown in Table 1. Crushed basalt with size ranged from 2.36 to 12.5 mm was used as coarse aggregate. The measurement of bulk specific gravity (Gsb), surface-saturated-dry (SSD) bulk specific gravity (Gsb,SSD), apparent specific density (Gsa) and the water absorption on coarse aggregates were performed in accord-dance with ASTM C127-88 as shown in Table 2. The grain size distribution and other physical properties of coarse aggregate are shown in Fig. 1 and Table 2, respectively. Two types of natural sands, namely desert sand (DS) and river sand (RS), were used as fine aggregates. The selected DS was referred to as red dune sand from central Australia as previously mentioned and the NS was commonly-used sand from local supplier for casting reference concrete. The measurement of bulk specific gravity, SSD bulk specified gravity, apparent specific density and the water absorption on fine aggregates were performed in accordance with ASTM C128-12 as shown in Table 2. Sieving analysis was performed in accordance with ASTM C136-06 and the grading curves, fineness modulus, coefficient of uniformity and average grain size were determined as shown Fig. 1 and Table 2 respectively. The surface area of fine aggregates was determined by B.E.T. Nitrogen Adsorption technique in accordance with ASTM D5604 and the results are presented in Table 2. The chemical compositions of dune sand and river sand were analysed by X-ray Fluorescence (XRF) as shown in Table 1.

Table 1 Chemical composition of cement and fine aggregates. Constituent

Cement (%)

DS (%)

NS (%)

SiO2 Al2O3 Fe2O3 K2O Na2O CaO MgO SO3 Cl Loss on ignition

19.9 4.7 3.4 0.5 0.2 63.9 1.3 2.6 – 3.0

94.8 2.00 0.66 0.34 0.06 0.23 0.11 – – 0.83

96.7 1.05 0.56 0.12 0.12 0.08 – – – 0.29

F.J. Luo et al. / Construction and Building Materials 47 (2013) 131–137

for reference purpose. Water–cement ratio (w/c) was controlled at 0.5 by weight. All aggregates were used in their saturated surface dry condition and all concrete specimens were cast without admixture.

Table 2 Physical properties of aggregates. Aggregate properties

Coarse aggregate

Bulk specific gravity SSD bulk specific gravity Apparent specific gravity Water absorption (%) Fineness modulus Coefficient of uniformity Average grain size (mm) B.E.T. surface area (m2/kg)

2.820 2.883 2.910 1.20 – – 10.810 –

133

Fine aggregate DS

NS

2.456 2.581 2.720 3.92 1.009 2.131 0.244 280

2.428 2.540 2.652 3.44 2.158 2.885 0.710 960

2.3. Casting and testing The raw materials were mixed by a conventional constant-speed mixer in accordance with ASTM C192. Slump was measured immediately after the mixing in accordance with ASTM C172. Air content (by pressure method) and specified gravity of freshly-mixed concrete was determined using concrete air meter in accordance with ASTM C231. For each mix, a number of cylinders with 100-mm diameter and 200-mm height were cast in pre-oiled steel moulds. Specimens were demoded after 24 h and cured in lime bath. The mechanical properties of hardened concrete including uniaxial compressive strengths, elastic modulus and tensile splitting strength were determined at the age of 28 days in accordance with ASTM C39, ASTM C469 and ASTM C496, respectively. Results were taken as the average of three measurements.

3. Results and discussion 3.1. Workability

Fig. 1. Grading curves of fine and coarse aggregate.

As evident from Tables 1 and 2, DS and NS have similar chemical compositions and physical properties. The major differences between these two types of sands are in their gradation and grain sizes (Fig. 1). The NS has a continuous grading with Coefficient of uniformity (Cu) of 2.88 while DS has a more uniform gradation with Cu of 2.13. The DS is superfine sand with fineness modulus of 1.01, which is more than twice as much as for NS. Also, DS does not meet the grading requirement for fine aggregate specified in ASTM C33 and AS 2758.1. Moreover, DS has an average and maximum grain size of 0.244 mm and 1.18 mm, respectively, while for NS, they were 0.710 mm and 4.75 mm, respectively. Furthermore, although there is no contamination of dune sand with chlorides or sulphates, organic impurity due to decayed vegetation exists which makes LOI in dune sand slightly higher than in river sand as seen in Table 1. This may affect the hydration reaction as well as durability of concrete and requires further research.

2.2. Mix proportions The amount of VPF in mixture was manipulated by varying sand–aggregate (S/A) ratios and aggregate–cement ratios (A/C) in three different levels respectively, i.e. S/A = 0.23, 0.3, 0.42 and A/C = 3.94, 4.69, 5.44. In such way, a total of nine concrete mixtures were designed as shown in Table 3. The resulting sand–cement (S/ C) ratios are accordingly ranged from 0.91 to 2.28 which cover the typical range for normal purpose concrete; wherein the lower level of S/C represent a rich practical mixture and the higher level represents a lean mixture used for concrete filling. In this study, RSC samples with identical mix proportions were also cast and tested

Slump test was performed to evaluate the workability of fresh concrete. The measured slump is then plotted against its sand– cement (S/C) ratio as shown in Fig. 2. In the current research, as a constant of water–cement ratio of 0.5 was used for all mixtures, the higher the S/C ratio is, the higher and the sand–water ratio (S/W) will be. The corresponding S/W ratio is also labelled in Fig. 2 (see the top horizontal axis). Furthermore, the data is grouped accordingly to sand–aggregates (S/A) ratios which represent different aggregate skeleton structures. The results show a general trend that, for a given S/A ratio, the slump decrease with increasing S/C ratio (or decreasing amount of mixing water in mixture), regardless of the type of fine aggregate. It is noted that, for both DSC and RSC, the highest slump was observed in the mixture having S/C ratio of 1.18 (or S/W ratio of 2.36) with S/A ratio of 0.30, which was not the one having highest water content. The possible reason for this is described as follow. According to Wang et al. [18] the mixing water exists in the forms of free layer water, adsorpted layer water and filling water. The different forms of water play different roles in contribution to workability. Free layer water makes the solid particles separate each other and thus contributes to workability. The adsorpted layer water is very close to the surface of solid. Due to the adsorption of the solid surface to water molecule, this part of water will be restrained by solid particle and is not able to move freely. As a result, this water makes no contribution to workability. The filling water only fills into space among solid particles and makes

Table 3 Mix proportions. Code

S/A

A/C

S/C

No. No. No. No. No. No. No. No. No.

0.23 0.23 0.23 0.30 0.30 0.30 0.42 0.42 0.42

3.94 4.69 5.44 3.94 4.69 5.44 3.94 4.69 5.44

0.91 1.08 1.25 1.18 1.41 1.63 1.65 1.97 2.28

1 2 3 4 5 6 7 8 9

Density (kg/m3) DSC

RSC

2539.2 2565.4 2614.4 2535.9 2555.0 2552.3 2477.1 2493.5 2539.2

2526.1 2568.6 2594.8 2545.8 2590.0 2581.7 2493.5 2535.9 2539.2

Fig. 2. Slump versus sand/water ratio.

134

F.J. Luo et al. / Construction and Building Materials 47 (2013) 131–137

no contribution to the workability as well. It is noted that different forms of water can be exchanged in the system, which is dependent on the nature and the proportion of different solid particles. For example, when the proper proportion of fine and coarse particles is mixed together, the fine particles will fill the porosities of the granular structure (cement and aggregates) by releasing the water in these pores. In this way, the workability will be improved as part of filling water is transferred to free layer water. This may explain why the highest workability was observed in mixes with an intermediate S/A ratio of 0.3 rather than mixes with highest mixing water content. When the sand content increases, the excess amount of fine particles which cannot fill in porosities will absorb water on their surface. The amount of adsorpted layer water is dependent on the surface area of sand. As dune sand exhibits significantly higher surface area than river sand, dune sand particles have high amount of adsorpted layer water. Therefore, at a high level of S/W ratio (S/W > 2.8), the workability of DSC becomes significantly lower than that of RSC. The results suggest that a certain amount of mixing water is required to ensure DSC a comparable workability to RSC. Moreover, the S/A ratio plays important role in optimising workability of DSC when enough mixing water was presented (S/W < 2.8).

3.2. Air content and density In addition to slump testing, the density and air content were also measured for fresh concrete. The measured density for both DSC and RSC is given in Table 3. As presented in the table, the highest density among all samples is 2590.0 kg/m3 observed in RSC with an S/C ratio of 1.41; while the lowest is 2477.1 kg/m3observed in DSC with an S/C ratio of 1.65. The difference between the extreme values is less than 4%, indicating that the ratio of S/C has no significant influence on the density of both DSC and RSC. Moreover, at identical mix proportion, the difference between DSC and RSC samples is generally not greater than 1.5% which is negligible (see Table 3). The measured air content against its S/C ratio is presented in Fig. 3 for both RSC and DSC samples. As shown in the figure, there is a threshold level of S/C ratio above which the air content in DSC becomes more sensitive to S/C ratio At low level of S/C ratio (S/C < 1.41), the change in air content with S/C ratio is not significant; while at high level of S/C ratio (S/C > 1.41), the air content of DSC increases rapidly with S/C ratio, reaching a value of 1.7% at S/C ratio of around 2.0. In comparison to RSC, the air content of DSC is generally comparable at low level of S/C ratio while is notably higher at high level of S/C ratio. This trend is similar to the comparison of workability between DSC and RSC with respect to S/C ratio. This consistency implies that more air bubbles are introduced (during mixing process) in less workable mixtures. Although at this level of S/C ratio, the air content in DSC is relatively higher than RSC, it does not go beyond the typical range

Fig. 3. Air content versus sand/cement ratio.

for RSC. This range, according to Daniel and Lobo [19], is from 0.5% to 2.0% for entrapped air during mixing process. 3.3. Compressive strength The effect of changing S/C ratio on 28-day compressive strength (fc0 ) is shown in Fig. 4. In this figure, the solid line represents the compressive strength of DSC in relation to its S/C ratio. As presented, when S/C ratio is in low level, the compressive strength increase with S/C ratio, reaching a maximum of 59.6 MPa when S/C = 1.41. When S/C ratio is higher than 1.41, the strength decreases with increasing S/C ratio. The tendency implies that different mechanisms affect the strength of DSC in different levels of S/C ratios. This is further confirmed by a comparison of strength difference between DSC and RSC. The strength difference is shown in the bar chart in the same figure; wherein, the positive value indicates DSC having higher fc0 than its reference sample, or vice versa. It is evident from the bar chart that S/C ratio of 1.41 is also a threshold for the strength difference between DSC and RSC. When S/C ratio is under the threshold, the strength of DSC is comparable or even higher than that of DSC (both 4.08% and 4.26% are lower than the maximum standard deviation for compressive strength results). When S/C ratio is above the threshold, the strength of DSC becomes notably lower than its reference RSC. In this level, the strength difference may be attributed to the entrapment of more air bubbles in fresh DSC than in RSC (see Fig. 3). These bubbles occupy space between the cement grains and eventually create a more porous cement paste, which leads to the reduction in concrete strength. However, it is noted that the strength difference cannot be solely explained by the difference in air contents especially when S/C ratio is in a low level. For example, at S/C ratio of 1.08, the air content of DSC is slight higher than RSC whereas the strength of DSC is also higher

Table 4 Replacement rates corresponding to C/S ratios for different reference grain size. Code

S/C

p (dref = 215 lm) (%)

p (dref = 175 lm) (%)

No. No. No. No. No. No. No. No. No.

0.91 1.08 1.25 1.18 1.41 1.63 1.65 1.97 2.28

43 47 51 49 54 57 58 62 65

21 24 27 26 30 33 33 37 41

1 2 3 4 5 6 7 8 9

Fig. 4. 28-day compressive strength versus sand/cement ratio.

F.J. Luo et al. / Construction and Building Materials 47 (2013) 131–137

135

than RSC. This suggests that there must be other mechanisms for strength gain of concrete due to incorporation of dune sand. As previously discussed, the presence of VFP in dune sand can bring about strength gain due to heterogeneous nucleation and pozzolanic effect. Therefore, when studying the strength difference caused by using dune sand, one must consider the positive effect as well. To investigate the positive effect, the empirical model developed in Refs. [12–15] is applied as described by the following equations:

Df ¼

a 1 þ ð Sb Þ c eff

where Df is the overall strength gain due to heterogeneous nucleation and pozzolanic effect; at the age of 28 days; a is a time dependent coefficient and characterises the interaction between the cement and the VFP in dune sand which is equivalent to mineral admixture; b (=280) is a constant which close to the specific surface area of cement; c equals to 1 and; Seff is the efficient area defined by:

Seff ¼ S  nðpÞ ¼ Ss



 p  nðpÞ ¼ Ss  gðpÞ 100%  p

where Ss is the overall specific surface area of VFP in dune sand determined by B.E.T method (7077 kg/m2 for dref = 175 lm; 5890 kg/m2 for dref = 215 lm); S is the total area of contact; n(p) is the efficiency function as defined in the following equation and; g(p) is the efficiency factor

 nðpÞ ¼

1 þ cosðp  pÞ 2

k   p n 1  1þ m

where p is the replacement rate; k, m and n are constant which equal to 0.7, 36.8 and 3.40 respectively as suggested in the literature. To apply this model, it is required to define the fraction of VFPs according to a reference grain size (dref) above which the particles do not have significant influence on hydration process. With the fraction of VFP determined, the replacement rate (p) is then calculated as

p¼

VEP VEP þ Cement

The calculated replacement rates for DSC are given in Table 4. Here, it is worth mentioning that, although VFPs also exist in river sand, the proportion is considerably lower than (nearly half of) that in dune sand, equivalent to a replacement rate being close to zero. Therefore, in this model, only the VFPs in dune sand are considered. The dref, as proposed in the literature, is 215 lm; however, as this value is for ordinary mineral admixture other than dune sand powder, it may be required to define a new dref. In the current investigation, 175 lm was also selected as a possible dref. To evaluate these two dref, the numerical values obtained from the empirical model were compared with the experimental values. It was found that 175 lm is applicable for dref of dune sand and the detailed discussion is presented as follows. When 215 lm is adopted as dref, it is found that the strength gain decrease with increasing S/C ratio (see Fig. 5), which is not completely consistent with the strength difference between DSC and RSC as shown in the bar chart in Fig. 4). Accordingly to the bar chart, there should be a peak in the strength gain at S/C ratio of around 1.41, which leads to the highest positive strength difference between DSC and RSC. This is due to the fact that the positive strength difference can only be attributed to heterogeneous nucleation and pozzolanic effect because DSC and RSC have identical mix proportion, except for the type of fine aggregate. With dref adjusted to 175 lm, a peak in strength gain was observed in DSC at an S/C ratio of 1.41 (see Fig. 5), which matches better to the strength difference as shown in bar chart in Fig. 4. A further

Fig. 5. Tendency of strength gain due to heterogeneous nucleation and pozzolanic effect according to empirical model developed in Ref. [15].

examination of the strength gain (dref = 175 lm) shows an upward trend followed by a gradual drop-down at a certain threshold. This tendency is also consistent with previous literature. As pointed out by Cyr et al. [15], the efficiency of VFPs is different regarding strength gain when small or large quantities of powders are used: a small or moderate amount of powders exhibits increasing strength gain efficiency with increasing VFPs content while a large amount exhibits the opposite trend. Similar observation is reposted by Guettala [14] that the use of very fine dune sand powder as replacement for cement has positive effect on mortar strength when the replacement rate is lower than 20%, despite the fact that the type of dune sand is different. Therefore, 175 lm is more properly defined as dref for dune sand. Besides positive effects, the presence of VFPs can also have negative effect on concrete strength due to the entrapment of relatively large amount of air bubble (discussed above), and therefore, the final compressive strength of DSC is a results of their combined effect as shown in Fig. 6. Wherein, the bar chart represents the strength difference as it is in Fig. 4, the solid line represents the strength gain from heterogeneous nucleation and pozzolanic effect for DSC according to the empirical model and, the arrow represents the strength difference caused by different air content. The downward arrows indicate DSC having higher air content which relates to a negative effect due to higher porosity in concrete, or vice versa. At S/C ratio of 1.25, since there is no difference in air content between DSC and RSC, no arrow is labelled at this S/C ratio (see Fig. 6). As presented in the solid line, the beneficial effects are dependent on S/C ratio of DSC. When S/C ratio is lower than 1.41 (low level of S/C ratio), the beneficial effect increases with S/C ratio. Meanwhile, the detrimental effect due to the presence of more air bubbles counteracts the beneficial effects, leading to a strength reduction in DSC. This is evident in DSC at S/C ratios of 1.18 and 1.41, with which DSC exhibits a lower strength compared to their

Fig. 6. Schematic diagram showing combined effect on the strength of DSC.

136

F.J. Luo et al. / Construction and Building Materials 47 (2013) 131–137

presented in Fig. 8, DSC has comparable elastic modulus to RSC, and for both type of concrete, Ec increase with increasing fc0 . As suggested in ACI 318-02 (solid line in Fig. 8), the expression for secant modulus of elasticity of concrete, for structural calculation, applicable to normal weight concrete, is Ec ¼ 4:73ðfc0 Þ0:5 based on standard test cylinders. As suggested in AS 3600-2009 (dash line in Fig. 8), for concrete with mean compressive strength over 40 MPa, the elastic modulus of concrete at 28 days shall be taken as, Ec ¼ q1:5  ½0:024ðfc0 Þ0:5 þ 0:12, in which the concrete density q is taken into consideration; note that in Fig. 8, the density is normalised to 2547 kg/m3 which is the average value for all the specimens. Fig. 7. Tensile splitting strength versus 28-day compressive strength.

4. Conclusion The following conclusions are drawn from the study:

Fig. 8. Elastic modulus versus 28-day compressive strength.

RSC counterparts. For other samples at low S/C level, beneficial effects are dominated mechanism, leading to a positive strength difference between DSC and RSC. When S/C ratio is higher than 1.41, the beneficial effects decrease with increasing S/C ratio whilst the detrimental effect surpasses the beneficial effects, leading to an overall deterioration in DSC strength. As a result, the strength of DSC is lower than that of RSC. Note that the figures above (Figs. 5 and 6) are only schematic diagrams for predicting the tendency only, which does not indicate the magnitude of the effects. This is due to the fact that (i) the empirical coefficients in the model are developed based on mortar mixtures rather than concrete mixtures and (ii) heterogeneous nucleation and pozzolanic effects are dependent on the nature of fine powders; in this regards, the dune sand is different from the traditional admixtures. Therefore, a further research is required to develop the model to quantify the positive effect on strength of dune sand concrete due to presence of VFP. 3.4. Splitting tensile strength In the study, the measured splitting tensile strengths (ft) against their corresponding fc0 are shown in Fig. 7 for both RSC and DSC. For comparative purpose, different equations for estimating tensile strength are also shown in Fig. 7. As proposed in ACI 318, the expression ft ¼ 0:3ðfc0 Þ2=3 indicates the relationship between the splitting tensile strength and compressive strength based on standard cylinder specimens, as shown in Fig. 7 (solid line). In AS3600, the relation is predicted as ft ¼ 0:56ðfc0 Þ1=2 , shown in the same figure (dash line). It can be seen from the figure that the measured values for both DSC and RSC are above the expected values suggested by both ACI 318 and AS3600.

(1) Compared with reference RSC samples, the use of dune sand does not have negative effect on the workability of concrete up to an S/C ratio of 1.41, below which DSC has comparable slump to RSC. At S/C ratio of 1.18, DSC has its highest slump, reaching a value of 105 mm, which is only 5 mm lower than that of RSC. With further increase in the ratio of S/C, the slump of DSC becomes significantly lower than RSC as excessive VFPs absorb large quantities of water on their surface. (2) At low level of S/C ratio (S/C < 1.41), the variation in air content with S/C is negligible for both DSC and RSC. At high level of S/C ratio, air content increases with increasing S/C ratio and the growth is more notable in DSC. (3) Based on empirical model and experimental data, the reference grain size for VFP is estimated to be 175 lm below which the dune sand powders facilitate hydration process due to heterogeneous nucleation and pozzolanic effect. (4) With reference to RSC, DSC possesses a considerably higher amount of VFPs. The compressive strength of DSC is influenced by the following two mechanisms related to VFPs, namely (1) heterogeneous nucleation and pozzolanic effects which lead to strength gain and (2) the entrapment of more air bubbles during compaction (especially at high S/C ratio) which leads to strength reduction. The strength difference between DSC and RSC is a result of the combined effect from the above mechanisms: at low level of S/C ratio (S/C < 1.41), the positive effects are predominated making DSC comparable or even higher than RSC in fc0 ; at high level of S/C ratio (S/C > 1.41), the negative effect will surpass the positive effect, which lead to notable strength reduction in DSC. (5) A comparative measurement of tensile splitting strength and elastic modulus between DSC and RSC was conducted and the results show a similar performance in these properties with respect to a given compressive strength. (6) Results show that when the level of S/C ratio is properly controlled, dune sand can be used as a complete replacement for ordinary fine aggregate without negative effects on the engineering properties of concrete. (7) It should be noted that, however, the test trends observed are valid only for the type of desert sand studied. For the dune sand from other deserts, the optimisation of VFP content in DSC should be based on trial mixes.

Acknowledgments 3.5. Elastic modulus Fig. 8 shows the relation between elastic modulus (Ec) and 28-day compressive strength (fc0 ) for both DSC and RSC. As

The authors are grateful for the financial support provided by the Australian Research Council to conduct this study. Dr. He thanks Deakin University for funding her Alfred Deakin Fellowship.

F.J. Luo et al. / Construction and Building Materials 47 (2013) 131–137

References [1] Harrison DJ, Steadman EJ. Alternative source of aggregates, B.G.S.C. Report, Editor; 2003. [2] Macfarlane M, Mitchell P. Scoping and Assessment of the environmental and social impacts of river mining in Jamaica, MERN Working Paper No. 322003: University of Warwick. [3] Pilkey, O.H., et al. Mining of coastal sand: a critical environmental and economic problem for Morocco. [4] Harthy Al, A.S., et al. The properties of concrete made with fine dune sand. Constr Build Mater 2007;21:1803–8. [5] Zhang G et al. Performance of mortar and concrete made with a fine aggregate of desert sand. Build Environ 2006;41:1478–81. [6] Jin BH, Song JX, Liu HF. Engineering characteristics of concrete made of desert sand from Maowusu Sandy Land. Appl Mech Mater 2012;174–177:604–7. [7] Wang WH, Han LH, Qing SE, Ren GP. Tests on load sectional capacity of fine aggregate of desert sand concrete filled steel tubular stub columns. Eng Mech 2011;26:320–8. [8] Qin SY, Wang WH, Han LH, Jia YH. Tests on flexural strength of fine aggregate of desert sand concrete filled steel tubular rectangular beams. Eng Mech 2012;27:250–60. [9] Khay SEE, Neji J, Loulizi A. Compacted dune sand concrete for pavement applications. Proc ICE – Constr Mater 2011;164(2):87–93. [10] Sanjayan G, Stocks LJ. Spalling of high-strength silica fume concrete in fire. ACI Mater J 1993;90:170–3.

137

[11] Han MC, Han CC, Lee GC. Mixture and material factors affecting the strength and shrinkage of high performance concrete. J Asian Arch Build Eng 2006;5:145–51. [12] Lawrence P, Cyr M, Ringot E. Mineral admixtures in mortars: effect of inert materials on short-term hydration. Cem Concr Res 2003;33(12):1939–47. [13] Cyr M, Lawrence P, Ringot E. Mineral admixtures in mortars: quantification of the physical effects of inert materials on short-term hydration. Cem Concr Res 2005;35(4):719–30. [14] Lawrence P, Cyr M, Ringot E. Mineral admixtures in mortars effect of type, amount and fineness of fine constituents on compressive strength. Cem Concr Res 2005;35(6):1092–105. [15] Cyr M, Lawrence P, Ringot E. Efficiency of mineral admixtures in mortars: Quantification of the physical and chemical effects of fine admixtures in relation with compressive strength. Cem Concr Res 2006;36(2):264–77. [16] Guettala S, Mezghiche B. Compressive strength and hydration with age of cement pastes containing dune sand powder. Constr Build Mater 2011;25(3):1263–9. [17] Alhozaimy A et al. Properties of high strength concrete using white and dune sands under normal and autoclaved curing. Constr Build Mater 2012;27(1):218–22. [18] Wang A, Zhang C, Sun W. Fly ash effects: I. The morphological effect of fly ash. Cem Concr Res 2003;33(12):2023–9. [19] Daniel DG, Lobo CL. User’s guide to ASTM specification C94 on ready-mixed concrete. ASTM Manual Series. ASTM International; 2005.